Assessing the Economic Advantage of a College Degree

Our aim in posting to this blog is not to do ideological battle, but to advance the practice of grounding policy discussions in reliable evidence. It isn’t always easy to resist fighting over conclusions instead of sticking to assessing the quality of evidence and argument. People like us who have devoted considerable time and energy to analyzing educational opportunities, structures, and policies have developed strong opinions about the problems we see and potential solutions. It takes quite compelling evidence to induce us to modify our positions. Moreover, it is not so easy to determine which evidence is most reliable.

For example, simple descriptions of trends may be totally turned on their heads by a small change in the starting year of a data series. The gap between the immediate college-enrollment rates of high-school graduates from the highest- and the lowest-income families has declined over the past decade from 31 percentage points in 1998 to 26 percentage points in 2008. That is encouraging news. A positive trend. But in 2002 the gap was only 22 percentage points. So, did the gap plunge and then skyrocket—or was 2002 a blip? Moreover, it is not clear whether we should be looking at the gap in terms of percentage points or proportional differences. These two approaches frequently tell quite different stories.

If data-savvy economists have trouble sorting through these issues, the general public has little choice but to take what they read at face value. An object lesson in taking what one reads at face value was provided by our colleague Richard Vedder, in his recent post on this site entitled  “The Diminishing Economic Advantage of a College Degree.”

Dr. Vedder reports that the unemployment rate of college graduates was a third that of the general population in 1970 and is only about half now. Vedder takes this as one indication that the college degree has “declining value”.   But it is not meaningful to compare college graduates to a group (the general population) that includes college graduates. Vedder overlooks the fact that because college graduates make up a larger share of the population now than in 1970, their unemployment rate now has a greater impact on the overall unemployment rate. The only reasonable way to make the comparison is to compare the unemployment rate for college graduates to the rate for others who have not completed college. The rate for bachelor’s degree recipients was 51 percent of the rate for high school graduates in 1970, and 48 percent in 2009 (and has fluctuated over the intervening years).  

If this point isn’t clear to you, think of a simple analogy. Suppose the number of people driving hybrid cars with excellent gas mileage rose rapidly as a share of the population. Obviously the average miles per gallon of American cars would go up, and the ratio of a hybrid’s mileage to the average car on the roads would fall. But equally obviously, that doesn’t reduce the mileage advantage of driving a hybrid compared to any particular other car, like an SUV.

While Vedder’s treatment of the employment advantage may simply reflect an honest confusion, it’s hard to be as charitable regarding his treatment of earnings differentials. He reports that mean annual pay of females with a bachelor’s degree was 79.4-percent higher than their high-school graduate counterparts in 2000, and declined to 76.7 percent in 2008. He does not trouble to report that in the immediately prior year, 2007, the differential was 84 percent and in the succeeding year 2009, it was 81 percent. He also doesn’t explain why he made his unemployment rate comparison using 1970 and the most recent data, while for earnings he looked at two years in the most recent decade. In 1970, the mean annual pay of male college graduates—whom Vedder ignores in his discussion—was 40-percent higher than that of high-school graduates; in 2009 it was 91-percent higher (the  largest gap since at least 1962). For women, the differential was 57 percent in 1970, compared to 81 percent last year.

As Daniel Patrick Moynihan once said, you are entitled to your own opinion but not to your own set of facts.

(Our thanks to Jenny Bendewald for prompt and skillful research assistance.)